Module 3 Assessment: First Supervised Baselines
Assessment ID: ML-M03-QA01 Estimated active time: 35-45 minutes Status: Draft
Part A: Concept checks
Answer in one or two sentences.
- What is the difference between regression and classification?
- Why do we split data into training and test rows before fitting?
- Why should the target column not be inside
X? - What does a dummy baseline tell us?
- What does MAE mean in the quiz-score example?
- Why is accuracy only a first classification metric?
- What is a predicted probability?
- Why does one train/test split not prove that a model is ready for real use?
Part B: Applied task
Use the supplied synthetic Module 3 datasets.
- Train a median dummy regressor and a linear regression model for
quiz_score_day10. - Report test MAE for both.
- Train a most-frequent dummy classifier and a logistic regression classifier for
completed_module1_by_day10. - Report test accuracy for both.
- Inspect at least five predicted probabilities from the classifier.
Part C: Explanation
Write 4-6 sentences explaining:
- which candidate beat its baseline;
- how to read the regression error;
- how to read the classification accuracy;
- one reason the classification metric is incomplete; and
- two limitations of the exercise.
Rubric
| Level | Evidence |
|---|---|
| Pass | Correctly separates regression and classification; excludes targets from features; uses train/test split; reports both dummy baselines and candidates; interprets MAE and accuracy in plain English; explains limitations. |
| Revise | Runs most of the modelling steps but misses one important explanation, baseline, or target/feature boundary. |
| Not yet | Uses the target as a feature, evaluates on training data only, omits baselines, or claims the model is ready for real learners. |
Safety rule
Do not use real personal, confidential, employer, client, health, financial, authentication, or sensitive data.